Lesson 2 of 6~2 min read

Lesson 1 — What Makes Something an Agent

This is the most important lesson in the course. Most people use the word "agent" to mean "AI that does stuff" — but the technical definition is more specific and more powerful.

Lesson 1 — What Makes Something an Agent

This is the most important lesson in the course. Most people use the word "agent" to mean "AI that does stuff" — but the technical definition is more specific and more powerful.

An agent is an AI that:

Receives a goal

Decides what actions to take to achieve it

Executes those actions using tools

Observes the results

Decides what to do next based on what it learned

Repeats until the goal is complete

This loop — **think → act → observe → think** — is called the **ReAct pattern** (Reasoning + Acting). It's the backbone of every serious AI agent built today.

**The key insight:** The AI is not just generating text. It's making decisions. It sees a goal, reasons about what information it needs, calls a tool to get that information, reads the result, and decides whether it has enough to answer or needs to keep going.

**What "tools" means:** A tool is any function the AI can call. Examples:

`search_web(query)` → returns a list of URLs and snippets

`read_page(url)` → returns the text content of a webpage

`write_file(filename, content)` → saves output to disk

`send_email(to, subject, body)` → sends an email

`run_sql(query)` → executes a database query

You define the tools. The AI decides when and how to use them. That separation — AI decides, code executes — is what makes agents safe and controllable.

← PREVIOUSWhat You'll Build NEXT →Lesson 2 — Setting Up Your Tools